fraud 2026-05-31 12 min read the underwriting desk

Tuning fraud filters for peptide stores

3-minute scan
  • Default gateway fraud rules are tuned for low-risk retail and will either bleed chargebacks or block good peptide orders.
  • Start with AVS + CVV hard rules, layer velocity caps, then add 3DS selectively on high-ticket and repeat-decline orders.
  • Measure every rule against both chargeback ratio and false-decline rate before you make it permanent.
On this page

    A peptide operator we talked to last quarter had a clean offer, decent margins, and a chargeback ratio creeping toward 1.1%. His gateway fraud rules were stock — exactly what the integration wizard set on day one. AVS was set to "advisory only," CVV mismatches were allowed through, and there was no velocity cap at all. One stolen-card ring found his checkout, ran 40 orders in an afternoon, and three weeks later he was staring at a reserve increase and a warning letter. None of it was necessary. The rules to stop it were already in his gateway. Nobody had turned them on.

    This is the most common avoidable problem we see on peptide books. Fraud filters are free, they ship with every gateway, and almost nobody tunes them for the vertical. Here is how to do it without strangling your conversion rate.

    Why peptide stores need custom filter tuning

    Default fraud rules assume a low-risk merchant: low average ticket, mainstream product, customers who shop from their home billing address. Peptide checkout breaks all three assumptions. Average tickets run higher. Customers buy from work, from a VPN, from a shipping address that is not their billing address. And peptide stores attract more card-testing traffic because the category is known to run permissive default settings.

    So you get punished twice. The stock rules let real fraud through because they are set to "advisory." And when you panic and tighten everything, the same rules block legitimate repeat buyers who happen to use a forwarding address. The goal is a middle path: hard rules where fraud signal is unambiguous, soft rules where it is noisy, and 3DS where the ticket justifies the friction.

    Before you touch anything, pull 90 days of orders and tag every chargeback by reason code. If most of your disputes are genuine "I never authorized this" fraud, filters help a lot. If most are friendly fraud ("I never received it" from someone who did), filters help less and your refund and fulfillment policy matters more. Know which problem you actually have first.

    Layer one: AVS and CVV

    Address Verification (AVS) and CVV checks are the cheapest fraud reduction you own. They run on the issuer side, cost nothing extra, and catch a meaningful slice of stolen-card attempts.

    For a peptide store, set CVV to a hard decline on mismatch. There is almost no legitimate reason a real cardholder enters the wrong CVV three times. AVS is noisier because forwarding addresses and gift orders produce legitimate mismatches, so set AVS to decline on a full no-match (N response) but allow a ZIP-only match (the common Z and A partial responses). That single configuration stops the lazy card-testers without blocking the customer shipping to a reshipper.

    SignalRecommended actionWhyFalse-decline risk
    CVV mismatchHard declineAlmost never legitimateVery low
    AVS full no-match (N)Decline or reviewStrong fraud signalLow
    AVS ZIP-only matchAllowCommon with forwardersMedium if blocked
    AVS unavailable (intl)Route to 3DSNo data, not proof of fraudHigh if blocked

    Layer two: velocity caps

    Velocity checks limit how many attempts a single fingerprint can make in a window. This is the single rule that would have stopped the card-testing afternoon described up top. A velocity check counts attempts by card number, by IP, by device, and by email, and blocks once a threshold trips.

    Sensible starting caps for a peptide store:

    • Max 3 card numbers per IP per hour.
    • Max 4 declined attempts per card per day, then a 24-hour lockout.
    • Max 2 distinct billing names per device in a session.
    • Max 5 orders per email per 24 hours (raise for genuine bulk buyers you know).

    Velocity rules almost never hurt conversion because real customers do not retry a dozen cards. Card testers do exactly that. Turn velocity on first, watch for a week, and only loosen if you see legitimate buyers tripping a cap.

    Layer three: selective 3D Secure

    3D Secure (3DS) adds an issuer authentication step and shifts most fraud liability to the issuer on authenticated transactions. The cost is friction — a small share of customers abandon at the bank challenge screen. So you do not put 3DS on every order. You put it where the risk justifies the friction.

    Trigger 3DS on: orders above a ticket threshold you set (commonly two to three times your average order value), any order where AVS came back unavailable, any first-time buyer using a mismatched billing and shipping country, and any card that has already tripped a velocity decline today. Let your clean repeat buyers through frictionless. This selective approach typically cuts fraud chargebacks hard while costing you only a point or two of conversion on the flagged slice, not your whole book.

    Layer four: blocklists and allowlists

    Every gateway lets you maintain manual lists. Use them. When you win or lose a chargeback, add the card fingerprint and email to a blocklist. When a high-value repeat customer trips a rule by accident, add them to an allowlist so they sail through next time. These lists are tedious to maintain and worth every minute — they are the only fraud control that learns from your specific buyers.

    One caution: do not blocklist by IP range alone. Peptide buyers use VPNs heavily for privacy reasons that have nothing to do with fraud. Blocking a VPN exit node blocks dozens of legitimate customers along with one bad actor.

    Measuring whether your tuning worked

    Two numbers tell you if a rule is helping: your chargeback ratio and your false-decline rate. A rule that cuts chargebacks but quietly declines good orders is not a win — you just moved the loss from disputes to abandoned carts where you cannot see it.

    Track both before and after every change. Keep your ratio comfortably under the card-network thresholds — Visa flags excessive at 0.9% and Mastercard at 1.0% — with real headroom, not right at the line. And watch your decline rate week over week. If declines jump after a tightening, you over-corrected. Loosen the noisiest rule, usually AVS, and re-measure.

    Where orchestration changes the math

    If you run a single peptide brand, all of this lives inside one gateway and you tune it in one dashboard. If you run several brands, fraud tuning gets harder fast: each gateway has its own rule engine, its own blocklist, and its own thresholds, and a card-tester banned on brand A walks straight into brand B. A parent merchant account with shared fraud rules across brands solves the cross-brand gap — one blocklist, one velocity engine, one place to tune. That is the model we orchestrate for multi-brand peptide operators. We do not process the payment ourselves; we sit on top of your acquirer and gateway so the fraud controls work consistently across the whole portfolio. For a single brand, tuning your existing gateway is the better call and you do not need us. See how the layers fit in how it works.

    If you are running three or more peptide brands and your fraud rules are fragmented across gateways, an honest fit check takes about twelve questions and gets you a straight answer on whether consolidating fraud controls under one parent account is worth it. If you are single-brand, we will tell you to keep tuning what you have.

    Found this useful? Share it X LinkedIn Reddit HN Email

    FAQ

    Will turning on strict fraud filters hurt my conversion rate?
    Velocity caps and CVV hard-decline almost never hurt conversion, because real customers do not retry many cards or fumble the CVV repeatedly. The rule that costs conversion is 3D Secure, which adds a bank challenge step. That is why you apply 3DS selectively — only on high-ticket orders, unavailable-AVS orders, and cards that already tripped a velocity decline — instead of on every checkout. Measure your false-decline rate after each change so you catch over-tightening before it quietly drains your revenue.
    What chargeback ratio should I be tuning toward?
    Aim for real headroom under the network limits, not the limits themselves. Visa flags excessive disputes at 0.9% and Mastercard at 1.0%, with severe tiers around 1.8 to 2.0%. Tuning to sit right at 0.85% leaves you no margin for a bad week. Target 0.5% or below as your operating goal so a single fraud burst or a slow representment cycle does not push you into a monitoring program. See our chargeback threshold entry for how the programs escalate.
    Should I use a third-party fraud tool or just the gateway rules?
    Start with the gateway rules. They are free, already integrated, and most peptide stores never exhaust them before tuning AVS, CVV, velocity, and selective 3DS. Third-party fraud platforms add device fingerprinting and machine-scored risk that help at higher volume, but they cost real money per transaction and add integration work. Get the free layers right first, measure your residual fraud, and only buy a paid tool if a specific gap remains after tuning everything the gateway already gives you.
    How do I stop card testing on my peptide checkout?
    Card testing is high-volume, low-value attempts to validate stolen card numbers, and velocity rules stop it cold. Cap card numbers per IP per hour, cap declined attempts per card per day with a lockout, and rate-limit your checkout endpoint itself. Add a CAPTCHA or invisible challenge after the second failed attempt from a session. The testers move on the moment your store stops being an easy validation target, and your gateway processing fees on all those failed attempts stop bleeding too.
    Does 3D Secure move chargeback liability away from me?
    On most successfully authenticated 3DS transactions, fraud-related chargeback liability shifts to the card issuer rather than you, the merchant. That is the main reason to route risky peptide orders through it. The shift applies to true fraud disputes, not to friendly fraud or product-not-received claims, which remain your problem and are better handled with tracking, delivery confirmation, and a clear refund policy. Treat 3DS as fraud-liability protection on the orders most likely to be genuinely fraudulent, not a cure-all.
    My filters are tuned but chargebacks are still climbing — what now?
    Re-tag your recent disputes by reason code. If they are genuine fraud, your filters have a gap and you tighten velocity or expand 3DS coverage. If they are friendly fraud or not-received claims, no filter fixes that, and the lever is your representment evidence and fulfillment proof. A climbing ratio with clean fraud filters almost always means a dispute-handling problem, not a fraud problem, and the two need completely different fixes.

    Running multiple brands?
    multiflow was built for this.

    The Operator Briefing

    Twice-monthly. No fluff.

    Processor shutdowns, reserve-hold playbooks, reconciliation lessons, and the merchant-account decisions that save operators six-figure years. Delivered to your inbox — never spam.

    No spam. Unsubscribe in one click.

    We use essential cookies · Privacy